Sunteți pe pagina 1din 16
Coalbed Methane: Current Field-Based Evaluation Methods E.R. Clarkson, Universty of Calgary, and RM. Bustin, University of British Columbia ‘Summary CCoalbed methane (CBM) produced from subsurface coal deposits thas been produced commercially for moze than 30 yeas in North ‘America, and relatively reeenty in Australi, Chine, and India Histocical challenges to predicing CBM-well performance and Jong-tesm production have included acurete estimation of g33 in place (including quantification of in-situ sorbed gas storage). estimation of initial fluid ssturations (in saturatod reservoirs) and ‘mobile water in place; estimation of the degree of undersatura- tion (andersaturated coals produce mainly water above desorption pressure); estimation of inital absolute permeability (eystem); ‘selection of appropriate relative permeability curves; estimation of bsolte-permeabilty changes a8 a function of depletion; predic~ tion of proguced-gas composition changes as a function of deple tion; accounting Tor mulilayer bebavior, and accurate prediction of cavity or hydulic-facture properti, These challenges have primatly been a result ofthe unique reservoir properties of CBM. ‘Much peogress has been made in the past decade to evaluate fundamental properties of coal reservoirs, but obtaining accurate ‘estimates of some basic reservoir and geomechanical propeties ‘remains challenging. ‘The porpose of the current work isto review the sate of the ant in feld-based techniques for CHM reservoie-property and stimulationofficiency evaluation. Advances in production and pressure-ransieat analysis, gos-conteat determination, and mate- ‘al-balance methods made in the past 2 decades will be summa rized, The impact ofthese new methods on the evaluation of key reservoir properties, such as sbsolutfrlative permet and ges ‘contetlgas in place, as well as completion/stimulation properties will be discussed. Recorumendations on key surveillance data {0 assist with Geld-based evaluation of CBM, along with insight into practical usage of these dats, will be provided, Introduction (CBM has grown to bea globally signiSeant natral-gas supply, yet ‘our undortanding of key reservoir properties continues to evolve. ‘The unique storage and transport properties associated with CBM reservoirs as required that conventional petroleum engincer- ing methods be modified significanily in order that quantitative analysis of field data ean be performed. In the early 1980s, the Gas Research Insitute (GRI) worked to develop technology for recovery of CBM inst culminated in the mid-1990s with th excellent pblications on CBM assessment and development, much of which remains relevant today. These volumes iacluded detailed descriptions of reservoir engineering methods for gas-content ‘and g2s-inplace determination (see McLennan etal (1995) and Mavor and Nelson (1997), respectively] as well as pressure ra- sient (well-test) and) materal-balanee analysis (see Mavor and Saulsberry (1996) and Zaber (1996), respectively] and reservoir Simoltion [see Paul (1996). In the mid-1990e, those publications (in wddition o important research performed by industry, gover= iment, und academia upto that poing) were of tremendous vale a5, {guide to optimal development of CM lays. Up to that, ony 24 acy of am Ermene ‘pre [SPE 131720) suo scoped fe pasa He SPE Ure Ss Stearn, tory, Pereyra UAB Frey 240, need bir Ofna rane maven Te Fru Sb1 Maced sore ‘sed co Papper ape xb Si ia o ‘however, CEM development was resticed primarily to continental US basins, of which only (wo, the Black Warioe and San Juan busins, had achieved significant commercial success, Since that time, CBM development is expanded significantly ia the US to other onshore basins, such as the Uinta and Powder River basins, and indeed globally with commercial CBM succestes in Canada and Australia and CBM exploration projeots in China, India, and Indonesia, to name only a few. Outside of North America aud ‘Australis, however, CBM exploration to date has led to litle or ‘no commercial development, even though substantial exploration has taken place and large volumes of gas have been proved to be presont. Faure of CRM prospects has invariably been caused by low permeability of the coals and zesuling subeconomie produc tion rates rather than low gas contents, Curent, asa result of the combination of comparatively low gts pices and the success of 215 shales, CBM exploration in mest parts ofthe Woeld hat wened, with a few notable exceptions. (CBM reservoir property and well-stimulation efficiency eval ation as improved in the past decade to keep pace with lesmings from new basin exploration and development activity (Nox ‘Amevicn and beyond), as well as advances in dillin, completion, ‘and stimulation technology. The objective ofthe curent paper sto review the corent state of the art in fied-based methods forevalas- tion, The focus ison advances made in several Key areas inthe past 15 years, since the publication of the GRE values on CBM in the rid-1990s. The paper is not meant tobe exhaustive, We consider those aspeots that we believe have had and will have the greatest impact on future CBM exploitation. A separate paper on advanced Taboratay-based methods wil follow in tho nedr future, Review of CBM Reservolr Characteristics, Coa is both a source rock and reservoir for gas, Because ofthe many wnighe properties ‘of col, some of whieh aré summarized inthe following, qoactly- ing reservoir properties and stimulation efficiencies and forecasting prodvetion rates af coal ate challenging andl equine unique analy fal tools and theoretical analyses. During eatly burial, biogenic 2s, and ducing later burst, dhermogenic gas, is generate, some Df which is retained inthe Coal in the sorbed state because of the high surfsce arew of the organic mater, Because the gas in the sorbed state has a comparatively high density, significant volumes ‘of gas can be stoved even at relatively low resevoir pressure. Compared to conventional reservoir rocks, coal has a vey low Young's modulus and high Poisson's ratio, and coal invariably ‘nas one or two sets of closely spaced extension fractares (cleat) formed during the nor) course of coalification and/or dering subsequent deformation. The solid mattix between eal fractures (cleat) is comparatively impecvious, andthe matrix swells signif: cantly with gas sorption. These properties combine to rele fn 8 reservoir in Which fraciure permeability may increase or deczease with production and in a reservoir that as propensity to have re-crsting fines andr one in which fines are produced dating tilling and completion. The primary mechanisms for gus storage ia CBM are (1) adsorption upon the internal surface wea associated with organic matter, (2) conventional (iree-gas) storage in natural fractures and in matrix porosity, an (3) solution in bitumen and formation water, Not that we ase the etm “sorption” to encompass adsomp- jon of gas on the intemal surface aoa of coal and solvation of ss by liquidsolids inthe coal maitix ~ when somption isotherms ‘re measured inthe laboratory for establishing ga conten, these Fetwuay 2011 SPE ReseroieBralustion & Engineeting 2-Phase Flow In Fracture (Barey) Fig. 1A conceptual mode! for the flow of gas and water {thtough coal. Note that single-phase flow of ges is thought {@ occur through the coal matrix by means of the mechanisin ‘of dituston, whereas wator and ges flow through the fracturs (eleat”) by Darcy flow. mechanisms of storage are typically not distinguished, ree-pas and solution gas capacity in costs has generally aot been considered in iantifying reservoir capacity Gr gas conten, However, sides by Bustin and Clarkson (1999) and Bustin and Bustin (2003) shows ‘that in low-rank coals in which pores are comparatively large aud surface areas comparatively small, free ges may form a signilie ‘cant componeal of the reservoir capacity and gus content. This is, particularly the ease in dry coals in which the pore and fracture systems are not saturaled with water, such a the dry Horseshoe (Canyon coals in Alberta, Frees eapacity of coals i determined ‘sing analytical methods aki those used in gas shales—namely, ‘core preservation a the wellste, measurement of pore volome, and ater saturation, Solution gas has similarly beon ignored ia mast coals. In low-rank coals euch of those of the Poder River basin of ‘Wyoming or Horseshoe Canyon coals of Alberta, however solution 83 can be shown to be» measureable and soonamcally important Bat of the resources (Bustin and Bustin 2009) Sorte gas storage is usually the most imporiant storage mech- anism in CBM reservcirs, owing tothe immense intemal surface see of coal, High-rank coals can have surfsce ares inthe range ‘of 100-300 mg, whereas conventional reservoirs typically have ‘afc areas < Tm, Most ofthe surface are ofthe coal natin 's associnted with onyanic meter, with porosity in the aacopore rnge. CBM matix pore soucturs is a inetion of thermal mat sity (rank), cnganimatier content, and composition, The controls pon sorption, in addition to the organic matter pore strvetre, Include: temperature, presure, moisture content, thermal atu nity, iineral matter Content (grade), organic matter composition, {nd gas composition. Sorption on coal bas been modeled with ‘arety of empirical 2d toeoretieal equations. The most commonly ‘applied gas-adsorption mode! for coal is the Langmuir (1916) isotherm, versions of which can be used for modeling single and !wulticomponcnt sorption. The Langmuir equation can be used 0 ‘state coal gas content if the coal seams are saturate (ce, the in-sito gos cootent is equal othe in-sity storage capacity), eservoir ‘ess, and ges composition ae known securaiely,fe=-gas and Februmy 2011 SPE Reservoir Eveluaion de Engineering solution gas storage are nealisibe, and the average composition ofthe reservoir is known, Flow through coal is generally considered to occu by different mechanisms, depending on whether flow is through the matrix or the fractures (Fig. 1). The primary mechanisms for gus flow in coals are considered to be (1) pressure-driven How through the fractures and (2) cancentration-deven low through the eoal tmatsix. Gas production inthe past generally has been considered limited by viscous or Darcy Now in fractares rather haa by mati siusion rates. Mavor (1995, p, 4.28), for example, suggets thet itis unlikely that reservoirs with conventional permeabiity sut- able for economic production would be hindered significantly by low diffusivity. In practice and theory however, whether or act ‘economic production is rate limited by mattix diffusion or by ‘coal-seam fracture permeability will €epend on effective fracture spicing, which determines diffusion low length, fracture perme ability, and diffusivity of the matrix (Cui and Bustin 2006; Bustin etal, 2008), Although most coals have very regular cleat neswork Fig. 1) commonly spaced at mich less than 1 em, it does not necessarily follow that these fractures have effective permeability, 28 supported by the extremely low permeability of maay coals ‘ith well-developed cleat networks sich a those ofthe Manville group in Alberta and the Camea cosle of Colorado Movement of gas out af the coal micropores and tensitiona ‘mesopores has long been considered controlled by ditfusion (ater than viscous Darey flow). Diffusion is generally modeled vsing Fick's law, which stsumes that mase-flow rate is proportional to the surfye ares and the diffusion coefficient of the material, Diffeson cceuts in the direction of lower coneantation. Diffusion is genet= ally considered in toms of three components (Cussler 1997) 1. Bulk diffusion—diffusion dominated. by intermolecalar Interactions and encompassing the diffusion of one species through diferent molecular species 2, Knudsen diffusion—difasion in which molecular snd pore= ‘wall interactions are important 5, Surface diffusion diffusion that oceus inthe adsorbed state fluid without mass transfer into the feo-gas state ‘The coal matrix provides @ source of gas othe natural fratues, If fracture spacing i fine enough, and/or te diffusion coefficient Js Tage enough, desorption may be asbumed to be insantaneous ‘and equilibrium models (Fig. 2) are adoquate. When fractures ere ‘more widely spaced andlor the ditfnsion coefficient is small then ‘desorption from the matrix ta the fractures isnot instantaneous, ard may need to be modeled using nonequilbrim approaches, ass ing one-stage or moltstge diffusion. Dual-poresity movels take into consdetation the movement of gas from the pritoary (matrix) porosity system to secondary (fractreleleat) porosity systems. These models consider the movement of gas fom the matsix to fimetures asthe source for he fractures. The dal-porosity formula ‘don of gas production of Waren and Root (1963), modified to take ino consideration matrix diffusion (Mavor 1996), is a8 follows: 4. =2.057V,9,01G,~G,), « here dye is gas-production rte from coal matsx (Macf/D), Vis volume OF coal matrix, i the eosl-matie shape factor (cr), D is the diffusion coetficient (emsec), G, 8 average gas content of the matrix (seton), and Gis storage capacity ofthe coal matrix evaluated at the average pressure of the natural-fracture system, (sefnon). The dual-porsity formation of diffusion is sill widely wed however, many recent studies have used 2 bidisperse (Ruckenstin st al, 1971) diffusion model far the coal matrix (as opposed to a unipore model), which is considered to more aptly represent the multiple pore sizes preseat in coul. The bidisperte models cur rently in use have bees summarized recently by Wei etal, (2007) ‘The bidisperse models assume Wo-step gas diffusion, with the micropore system dominated by surface diffesion and the meso and mactopore system dominated by pore diffusion (Fig. 2). In the bidisperse represeotation, gas sorption i8 in the micropores ‘and mesopores, and maeropores provide storage for free get anc 6 i I | | Egullorium ‘Approach No diffusion —+[ singlePorosty Simuiator [Clee porosity Nonequilibrium ff One-stage aftusion be ual Porosity Smmuistor Microporosty Cleat porosity i] ‘Two-stage diffusion eS) — ‘oth micro- ane ‘macropore diffusion { Mleroporosty \ eat porosty Fig. 2~Comparison of major elements of dit usion models applied to coal [moditod trom Wel etal. (2007)} provide the flow path to the fractures for sorbed gas stored in the micropores, ‘As a result ofthe interest in carbon dioxide sequestration in ‘coal seams, there has been a lege amount of research in the Inst vera yents esto the importance of diffusion, counterdffusion, 1nd adsorption kieties in mass transport through the eoal matrix, ‘The visualization is that carbon dioxide injetel ino the fracrurs network results in counterdiffusion and compeiitive adsorption n the micro- and mexopores, Elegant mathematical formulations such as those of Yi et al. (2008) suggest that cosiffusion of gas ‘molecnles facilitates gas wranspoit inthe presence of competitive ‘alsorplion whereas countesdifusion diminishes mass teanspor. Fracture flow is often modeled using some form of Darcy's lay, For two-phase CBM reserwirs, relative permeability effects hhave to be included and non-Darey flow may also be important. Tf coals are undersaturated, dewatering will be required before a {gs saturation develops in the fractures — in his cae, single-phase fiow of water will occur through the fractures until desorption pressure occurs, IF the coals are saturate, then two-phase flow of ‘5 will occur from the start of production. Absolute permeability in coal is highly dependent upon tho existence frequency, orienta tiga (clative to cureat in-situ stresses), fracture height and degree of mineral in-filling (Laubach etal. 1998). A common model for describing cleat porosity ancl permeshilty in coals the matchstick model (eidle 1992). The permeability is extremely seasitive to facture aperture, with which thas 2 cubic relationship. This model is simplistic becanse fracture apertures sve assumed t be uniform, nd pecmesbilty is assumed to be isotropic, Tn reality, fracture permeability may be anisocopic because of the greater continuity Office cleats celative to butt cleats ad because of the variation Of cleat apertures. Typical permeability anisowopies observed in coal are 2:1 to 4:1, and this is expected to have some impact on ‘development planning. Any process acting to modify the cleat aperture will have & strong effect on absolute permeability. In coul reservoirs, there ‘are two physical processes that will act to change the physical mension ofthe fracture apertures: (1) changes in effective stress and (2) mattix shrinkage. With Process 1, because the fracture pore volume is highly compressible, with pore-volume compress ibilities typically on the order of 10+ psi increases in effective swss because Of pore-peessure depletion Can cause the Fracture fapertures to decrease in wiih, which in tm exuses a redaction in absolute permeability. In some coal reservoirs, Process 2 will ‘ease absolute permeability to inrease with depletion, beeause the coal matrix will shrink during desorption, causing at increase in fracture apertures. Analytical models that ke into account both processes are reviewed in Palmer (2005), ‘Other important controls on fui flow through the fracture system include relative permesbilty effects (changes in effective o pennebility to gas and water during dewatering), reservoir pres suze and drawdowa, and fluid properties. Relative permeability is Inown to have a stong impact on CBM-preduction profiles. As two-phase CBM reservoirs are dewatered, the effective permeabil- 4ty to gas and water changes. The changes in effective permeability that ore fuid-satuation dependent are modeled using relative permeability curves, which ere notoriously dificult to obtain, For some CBM reservoirs, gos properties will change during deple tion not only because of a change in pressure bat also potentially because of ges-composition changes caused by relative adsorption effects. For example in the Fruitland coal fairway ofthe San Juan basin, where the gar in place is rich in CO, (10 mol% or more in some areas), because of higher sorption affinities of CO, excly Produced gas vas relatively enriched in CH, but dusing reser ‘oir depletio the relative proportion of CH, decreased and CO, increased to greater than 20%, ‘The following discussion focuses ou advances in feld-based techniques that have occurred over the past 15 years, since the GRI publications of Whe atid-1990s, and reflects application of few technology and leanings from nev basin development. The techniques have evolved 10 account for impraved knowledge of CBM reservoir properties and in the case of wellperformance analysis, to account for new wellbore architectures (horizontals, ‘mulilaterals, and mautifsetured horizontal wells) and stration Gax-Content and Gas-in-Place Determination. Gas in place for coal can bo determined volumetrically, using 2 combination of field ‘ata (eg, well logs and desorption canister tests) and Isboratony ta (Sue as sorption isotherms), or by means of materal-balanee tectniqoes Reservoir-Gas-Content Determination. The reservoir capacity ‘of coalbeds includes porosity available to free gas in pores and fractmes, sorbed gus principally on the organic fraction, and sol tion gas in bitumen and water. The gas conteat isthe smouat of in-place gas, which isthe sum of sorbed, solution, aud fre gas. If the gas content isthe same as the gas capacity, the coal is referred to as saturated, If less gas is present than the capacity, the col is referred to as unersatrated and a decline in reservoir pressure is required (0 optimize production ‘Alhongh 48 capecity has historically been determined in the laboratory through soeptionigtierm collection [see Mavor (1996), gas-content estimation still relies upon 2 combination of field and Inboratory data. ‘The waditional method is to perform desorption canister testing om multiple samples, then relate gis tontent to coal composition (specifically to inorganic fection) £0 Febvoary 2011 SPE Reservoir Blunion & Eaginesing that wireline logs (Le, density logs) may be calibrated to estimate the inorganic faction of coal in-situ and hence allow estenaton of _z85 content in-situ, This elaborate process, described by Mavor and ‘Nelson (1997), is necessary because, (o dete, there is n0 method avilable to directly deteet the ariount of in-situ sorbed gas, ‘There nas been some advancement in field-based gas-content {termination of coals during the fast 15 years, but forthe most, part these are only modifications of procedures established ithe 1976s. Ges content of coal is sill determined almost exclusively hy desorption testing using the tecimiques described by MeLen- nan etal. (1995). The principles of desorption testing were [aid ‘out initilly by Kissell etal. (1973) ofthe US Bureau of Mines (USBM) for measurements of gas content of coal. Subsequently, ‘minor modifeations have been recommended (australian Stancard 199%; MeLennan etal. 1995), overall, however, the techniques used today are essentially those of the SBM with some refinement Desorption testing is based on the premise that gas stored in cod] in the sored state takes considerable time ta desorb, and hence, if sample at the welste is sealed, subsequent ges eat is desorbed can be volimetically determined and used as & measure of the 2 in place in the reservoir Additional samples ean be elected or composition analyses, Samples for desorption analyses are col- lected ofthe wellsite and immediately placed i sealed canisters sid brought 10 the temperature ofthe drilling fluid or to « cal lated temperature depending on the sample and well conditions Desorption tests can be performed on conventions] coe, cuttings, or sigewall core, ‘The langestextors in desorption tests ere inthe estimate of gas {ost before sealing the sample (de lost gas") The amount of lost gus is dependent on the sample-retieval tie, the parile size (citfusivty), temperature, otal gas content, degree of saturation ‘of the sample, and natae of the dling fui (Ve air v3. water), ‘The mosteliable data are assured 10 be from wireline retieved core because of rapid cove recovery eompated 10 conventional core; the finer particle sive of euttings and sidewall core results in greater proportion of gas lost before sealing the samples in the canister: The lost-gas coroponent ean vary from § «9 >50% of the total gas content, and, heneo, iis measurement is critical or scourateg3s-n-place calculations and isthe largest source of error in gae-comtent dotermination. The theory and ality of the differ ‘ent methods of hast-gascalesation have becn seviewed by Mavor tal. (1994), Mavor (1996), Diamond and Schatzel (1998), and ‘Bustin (2005), Four different meshods are commonly applied 10 ‘measuring lost-24s volumes 2. USBM direct method 2. Smith and Williams 3 Amoco method 4.CBM solmions medhod The frst three methods ae described in detail by Mavor (1996) ‘nd will pot be reviewed here. ‘Th USEM direet method is the Most common method, end the Smith and Willams method is particule suite for well cuttings. Tre CBM solutions method is the newest method to appear and is deserbed briefly next ‘The CEM solutions mettiod (Bustin 2005) is base om the prem ‘se that ts impossible to determine the best analytical melt oF {£88 theoretical method for determining lost gas without know: jg shat the actual lost gas is. To develop a more robust lose gas ‘uethod, CEM solutions experimented with eutings, sidewall wore, comentionally reieve! cote, and wieline-retrieved care for both ‘ails and shale and developed specific sampling and lost gas pro © locos foreach sampling tebnique and lithology. Inthe laboratory iperiments, samples of different particle sie are saturated in the [aboratory and then desorption simulations are eared cut (Fig. 3). ‘Thus it possible io design a lost-gus procodute that fs the samples 4a assesses the magnitude of the potential eror in lost-gas eter ‘ination for varying lost-gas times and sample types During the eatly-time perio, the samples are typically nain- {eined at reservoir temperature oF mud temperature iy used for data follection during the lost-gas period. The reasoning is that diffe ‘ton tates are tempersiure dependent, However, hese temperatures fithcr over- or underestimate the actaal “average” temmperatare toy ofthe sample following coring snd sealing the samples in etry 2011 SPE Reservoir Braluation & Enginecting Under Estimation of Lost Gas ‘hour ‘Cube Contimetres 05 os 1 ts oe Square Root of Tine (hours) Fig. 3—Lost.gas simulation, Cuttings samples (0.099-0.192-in. equivalent diameter ar fuily saturates with gos, and then the sample is desorbed with decreasing pressure and temperature {0 simulate sample recovery during dling while desorbed Gas, | quantified, These data show that using the direct mothod of lastges analysis with @ t-hour lost-gas time, the error in Measurement is 41%, and if lost-gas time is'0°5 hours, the error ls 17%, =1 Pan ing pep, S507, Ake, * i PAP Derivation of tho equation assumes @ volumetric reservoir (00 sig nificant influx of water or change in reservoir volume), sotbed-gos storage only (no free-gas storage), no water sorage in secondary (Gracie) porosity, and that coals are saturated and sorption fol lows the Langmmir isotherm. Use of the equation, analogous t0 the conventional plz vs. cumulaive-proguetion plo, is iluseated in Hig 4 Errors in OGIP determination using this method could be caused day inaccurate pressure data; use of Langmuir pressures derived for Singl-component gases if the gas acwally cootsins multiple com ‘ponents; nonvolumetric reservoir bebavior; muldayer (ao crosiiow behavior) production in which high permeability zones deplete ‘aster than lower-permesbiliy zones, and existence of significant free-gus storage CGarkson and McGovern (2001, 2005) derived a simple exten- sion of the Jensen and Smith equation that includes fee-gas stor ge, which may be important in some CBM reservoirs (Bustin and Clarkson 1999; Bustin and Bustin 2009) p_ 22037405 Pte VBP 207355, fp. ib 0-S.)] 6) Ake” bpp WB am Although Bq, 3 allows for variation in water saturation during ‘epletion, Clarkson and McGovern assumed that this was constant for analysis, © avoid ieration—zhis assumption was justified carlier by Seidle (1999). Other assumptions in the derivation are similae to the Jonsen and Smith equation (Eq. 2) und Bg. 3 is used in similar way, except that the ordinate of the plot le-hand side of Eg, 3) requires knowledge of fre-gas-saturated porosity, [Langmuir volume, gas properties, and coal density. ‘Scidle (1999) developed a simplified version of the King (1990) equation, which isin p/c* format, where z* accounts for sorbed Bas storge: ® “a6, alike the more general King (1990) equation, a volumetic reservoir was assumed. Further, in practical usage of his equa- ‘ion, Seidie assumes! constant water saturation, which appears in the 2° definition, to avoid iteration, The use of Ba, is therefore analogous to the conventional p/z vs, curulative-produetion pot ‘xcept with 2° replacing z EUR OGip ‘Cumulative Gas Production Fig. 4—lustration of the use ofthe MMB equation for determi ration of original gas in place fand estimated ultimate recovery (EUR), given abendonment pressure). Periodically measured shut-in (reservoir) pressures are converted to dimonsiontoss form (P}P+?,) and plotted against cumulative gas production. Griginal gas in place (OGIP) Is obtained from a straight-line fit through the dimensionless pressure data and extrepolation to the sax, The intercept can be used fo cheek the initia proc ‘sure assumption, and slope can fe used to estimate drainag area (given coal thiekness, density, and Langmuir volume) oF ‘compare derived Langmuir volume (given drainage area, coal thickness, and density) to laboratory data. Finally, Ahmed et al. (2006) recently presented a new CBM materal-balance equation that accounts for wator-sturation changes during depletion and rock and water compressibility but, unlike Uae King (1990) method, doesnot require iteration. Beeause ‘ofthe complexity of the formation, the reader is eferel to the ‘original work forthe equations snd method of use. Given all of che mmateial-balance equations that are now available, iis logical to ask which one is appropriate for which reseevoir ype. In aur experience, given simulited, saturated-coal (Qwo-phase) coal-reservoir data that assumes a small amount of fee. os storage and no Water encroachment, ofthe methods just, fisted yield gts in-place and drainage-aea estimates well within enginesting ewor requirements (<1%). Additional simulation stud jes for dey CBM reservoirs with a higher amount of free-zas storage have demonstrated that drainage area calculated from the Tensen and Smith method start to diverge substantially from the input value. For Feld cases that we have analyzed (Satated, tWo- phase, low-free-gnsstorage, volumetric reservoirs), given the data ‘quality, the Jensen nnd Smith method yields reasonable results, and itrequires fewer reservoir data and isthe simplest of the equations tw implement. We leave itt the reader toy the various equations for thee date sets of inerest. Gas-Detiverability Determination. Evaluation and understanding, Of the doliverablity of gas from coal require quantitative assess: ‘ment ofboth matrix and coal-seam flow properties. Messarements snd modeling of coal-seam deliverability require a mixture of well tests and laboratory analyses, both of which have seen substantia refinement daring the last decade, The following wil focus only fn recent advancement of ficld-based techniques Measurement of Diffesion and Matrix Pransport. Toeortical rmadels of gas flux through the eatl matrix have advanced sap fly over the Ins few years [see Wei t al. (2007) for summary} however, there has beet limited progress in our ability to measure ‘matrix gas Flux, Commercial reservoir simulators forthe most part tse “sorption tine™ as a proxy for difsivity ofthe coal matrix (Mavor and Nelson 1997), Somption Gime (2) is defied as the time February 2011 SPE Reserva Bvaluton & Ensneoring required to desorb 63.2% of the gas from winole-core desorption tests Sonption time is related to che diffusivity, particle shape, and particle size ‘Most commonly, difusion is measured during canister desonp- tion tests, From the “direct method” of ost-ges ealelation (in ‘whieh desorption takes place at reservoir temperate), see) Gifusion coeticient (cm’isec), ? is characterise sion length (cm), G, is inital gas content (seffon), and 2, is the slope of the curve for desorption vs, square root of time {(settonyhr'. ‘The diffusivity can also be detived from the sorption time (Mavor 1996; King etal 1986): 0 @ Where AG, is cumulative gis desorbed (Sefton), ris ina, Gi ink tial gas conten, is sorption time (Seconds), snd ois coal-matrx shape factor (cm. The coal shape factors commonly used are 60 for cubes, 15 for spheres, and 8 for cylinder. ‘The diffasion coefficient of diffusivity determined from sorp- ‘ion time from canister tests essurnes thatthe temperature, dit sion Tengih, and diffusion rate of the eanister samples are similar 'o those of dhe reservoir and that desorption of gusos occurs at ‘constant pressure. Such assumptions, however, are probably rarely realistic “Mavor and Pratt (1996) applied more-sophistcated diffusion ‘models to desorption eanistor data fo the purpose of establishing ‘whether these morerigorous methods yielded more-accurte ost 8 and toal-gas estimates. The models investigated in that sy Included a bidisperse pore-iffusion made! ((vo-iffusion-coek- ficient model, assuming linear adkorption), models that included ar adsonption, and) a concentraion-dependent diffusion ‘model. The authors concluded that the mote-rigoroue modeling approaches did not yield moxe-aceurate estimates of Tost and {otal-gas estimates relative (othe simple direct method. Although scciracy of difusion-parameter esimates was not a focus ofthe sud, the authors mentioned thatthe bidisperse model provided sn advantage over the singlelifusion-parameter model because it was possible o match the entire desorption history as opposed to just eary.time data, ‘Recently, Cui etal 2008) presented a new method for deter- pemmeability ot diffusivity from desorption tests comdacted ‘on site. Those authors showed thatthe late-time desorbed fraction (4) ftom a canister desorption test ca be ited by in(t-¥,) Site ae 28) fom which the permeability can be determined as LE cscs) For earlystime data, the gas-desorption fraction can be epproxi- tated as eeavee tei we a ‘rom which the permeabitity can be detestnined as 4a BR Lds Oars? 6 any Fetmay 2011 SPE Reservoir Bvahtion & Enginesing Pe Derivation of Kgs. Bthrough 11 assumes thatthe length ofthe core is much lerger than the diameter: Altjough thie method is designed to extract permeabiliy, an effective gas diffusivity can be detived from the permeability through knowledge of measurement fluid properties and eoalsirativty. ‘Aa additional complication arises because af coal compressbil- assuming that diffusion at unconfined conditions in a canister isthe sane as coal confined in the subsurface is probably rarely if ever realistic, os suggested by permesblity measurements of | coal under varying confining pressures (Le, Bustin (1997), To Simulate these effects properly, laboratory methods such ax pulse ecay on confined cores at reservoir temperatures provide the most, realistic conditions for quantifying diffusion or permeabilcy of ‘coals however, methorsin common prectie use nomeservoir gases (Generally etiam or nitrogen) rather than reseevor gas (Broce fal, 1968; Trimmer eta 1980; Yared and Jones 1980; Hsich tal. 1981; Bourbie and Walls 1982; Dicker and Stas 1988; Jones 1997). Cui et al. (2000) discuss the failure of measurement of permeability and difusivity in tight rocks using custent methods ‘and provide new models for measurement of permeability endlor sifusion from pulse-deeay-type analyses of crushed end confined samples, from canister desorption, and from sorption isotherms. Bulk Reservoir or Fracture Absolate-Permeabilty Estimation Absolute penmeability in coal may be estimated feor laboratory core tests, wel-est analysis, advanced production-ata analysis, and from vesesvoir-simolation history matching. At present, we believe the most reliable and accurate method for establishing absolute permeability of coal is from carefully designed well tests, preferably performed during single-phase-fow (water or 4228) conditions, a6 can exist in undersaturated wells lowing above desorption pressure, or saturated reservoirs with nealgible rmobile-water production (dewatered coal wells or éry CBM res. cervoin). We will briefly discuss the recent advenoes in feld-based techniques below. 1. Permeability Derived From Pressure-Transient Analysis (P74), A variety of conventional well-est designs have been ured for CBM reservoirs including (prefractor and past fracture) injec= tion-flioff tests, (pot-ttec) flow/buildup tess, and tank and slog tests (Mavor and Stulsberry 1996). In adition to the examples provided by Mayor and Saulsbery, other examples of well-tst ‘applications that have appeared inthe SPE literature inthe past 20, ‘years (only select references are given) are given in Table 1. In recent years theoretical advancements have been made for analyzing before- and after-frctureclosure data associated with Injection tests in whict injection rates ate high enougl to initiate tnd propagate a hydraulic fracture, Tae new tesfmalysis proce- {ures circumvent some ofthe imitations of conventional well tet- analysis fechoigues applied to data in whic fractune presses have ‘been exceeded during the test, Recently, pretrectare. diagnostic injection tests (DFTT) have been used ia tight gas formations to obtain informtion for hydvauli-frcturing operations (leakotT ‘mechanism, closure time, and prestare) md resevoic pressure and permeability, the later vo coming fiom sftr-closure analysis, provided pseudolinear-and pseudoradial-flow vegimes are reached uring the test Barree etal. 2007), These tests have also been tried in coal reservoirs (Rauuethy et al, 2002). In Remurthy etal (2002), the possible Ieakoff mechanisms for CBM reservoirs were reviowed, and both before- and after closure analysis was applied to Canadian and US coal examples. The before-cosure analysis requires rock-property and hydrantic-racte data that are not ‘commonly available, and hence those results tend t0 be qual five. Further, itis no clear whether a hydraulic fracture is actually ‘eated in some coal DFTs, or whether the cleat sytem is simply inflated and Mui is propagated along an existing clea system with no new fracture created. The hydraulic-racture network may also ‘be complex see Palmer et al. (1993)], Finally, after closure analy- sis requires pseudaradial flow to develop, which may not ecco in reasonable ime frame for low-permeability coals. Tn addition to new theoretical developments, some advance- ‘meat in well-testimplementation technology hes also boen made For example, wireline-conveyed formation-evanation tools have been used recently (Schlachter 2007) to isolate coal intervals and 5 ‘TABLE 4—-SUMMARY OF CBM WELL-TEST STUDIES (SINCE 1990) “thor Toate Aralyesd Summary Zar eta PreraciFOT Discussed probioms with Seidl st al 4991 Pre-and postiracFOT time to roach radi! fv. ere nat discussed, Coals rom Mayor nd 1903 ST, FBU, Robison and ultivel interference Sut withthe single-wel tests. This work h basin Fruitland coe). Jochen and Lee 1993 DST. FBU, snd mulivot inteference ‘nalysis approach ylekied val shu etal 1904 Slugiprossure- uu packer tests ‘2 shaped Buldups that wore dieu to analyze. Conway etal «1895 —Shig, IFOT (gas and water), FB, Tank" malliphage flow on test rosuls was inves Drove reasonaole lowerirtostimat Sparksetal, 1995 IFOTISug covatal, 905 Fou Permeability was uncertain, traditional prorac imdation injection falloftest ierpretaton, inluding afte of exceeding formaiton pressure, testing muliple coal Seams, ad stess-depondence af pores and permeability New test procedures re proposed, Coals from the Black Warrior basin were analyzed, Discussed the effcte of tree gas and gas desorption on pre- end postracture stinlation infection falle.sost results, and deveiopad a consation for aquired shun 1 efets of stress-dependence of porosity and permeabilly ‘San Juan basin were analzee. ST, postcaviation (single-well) fomlbuldup tests and mutivelnterarence tess ‘were conlcted. Both “conventone” [using tho approach of Parina(1956)] and mol- ‘hase flow potental approachoe wero used and compared for postcaviation FEU, ‘where i wes observed tat tne ruliphase flow potential approach yielded results Consistent with reservar simulation. Mulwel interference tests yielded consistant re- ‘become the standard for muliphase flow CBM analysis The wolkests were performed atthe GR fst ste in the San Juan Simulated two-page CBM data, created for vurlousflowfshulin tines, was analyzed Using conventional anaiysie matnods and the modified pseucdpressure function and ‘approach suggested by Kamal and Six (1889). Nother approach yielded resus In oxsct ‘gresment with simulator inp, bul what was mero surpriing Was fal he conventional 5 doser than the Kemal and Six approach. As wih the Mavor and Robineon etudy, wall tests fom tho GRI test site wera performed, with Somowhat diferent estimates of absolute permoabily. Azone of altered (higher) permeability zone around cavtatod wells was inferred ‘Slug lestderved pemmeabilly and skin values were shown to be nonunique. Using a feld example fom the Rack Greek ale in tha Black Warroe basi, the authors ‘mention that sug teste perfornes by wihdrowing fuids should eonsier two-phase flow ‘fect, Prossure-buidup packer Less performed on core holes als coer displayed Several ypes of welests ware appt to the Mare Lee-Blue Creek coals atthe Rack {Greek prosicton pot stein the Black Warrior basin. Weltosts performed included si, ‘wales ene 903 Injection fot, produeionshutin, and tank twsis. The infence of ated, Wale Injection tests wore found to solute permeabily. The authors recom fended that if toting Is performed under multphase condiions, a tlowlong-term) Dullup test should be performed toyed gas- and wator-ffective permoabiliyinfor- ‘malin, followed by a water jection flo test to estat absolute permeatlty. They iso. suggest that feld estimates of rlaive permeabilty can be obtained analo~ {ously 10 procadvtes derived for eolulon gos-drve reservoirs (AbKhalie etal. 1987) Coat permeability wae quantiied a8 a function of stress—ato-tine fal ot data was, analyzed from IFOT tests in th Codar Cove ares, and slug tess wore analyzed in the Oak Grove area of th Black Warrior Basin. Minimum siress values in both areas wore oblalnod from sc-breskdlown tosis. Pormeatlily end production were demo- Petrated to correlate with stress. The authors rscommended performing IFOT tests ft 3s low rates 2s possi to reduce effects of stress. This shudy presents @ unique application of CBM waltast analyis—well testing was used to idenlly the source of matnane seopago observed inthe vicnly of CBM walls In southwestem Colored The wol tas Wars designed to obtain parmesbilly and reservoir presure Infomation for reservoir simuaton, skin frm producing wall and {o determine # pressure interference occured between preducng CEM wols and ‘nonlor wells, PBU fests were performed on monitor ara preducing Wolls. The Pectine (756) metod was used to calculate ffecive pormoabity to gas and wate, as wel fe skin. Permno's approach was. chosen over the ofullphas® pseudopressure fppreach because tha wols exhibited primarily sknle-phase Tow and relative ‘perform drawdowa/tuildup tests thar ean be analyzed using CBM PTA techniques. "There are several reservoir complenitis associated wth CBM. that serve to complicate design and interpretation of well texts: + Gas desorption (saturated CBM reservoirs or understurated reservoirs below desorption pressure) * Multiphase flow (atuated CBM reservoirs or understurated reservoirs below desomption pressure, two-phase flow of gas and wate) ‘* Matlayes bebavior + Additional reservoir heterogeneities: dual-poresity behavios, composite behavior ‘ Nonstati absolute permeability (dependence on effective stress andi sorption) ‘An excellent discussion of the theory associated with CBM ‘woll-tet analysis, and design and execution of single- and muli- well tests, was provided by Mavor und Saulsberry (1996)-—tech- higues to accoust for ens desorption, multiphase flow, and caulki- [ayer behavior were sarumarized. Recent studies have focused on teat designs for multilayered coals; moldlayer texting in single ebay 2011 SPE Reseroir Brabntion & Boginesing ‘TABLE {—SUMMARY OF CBM WELL-TEST STUDIES (SINCE 1980) (Continued) Yoor Tests Analyzed 1886 Interference Summary A muftinellinteserence test was used fo quantity permeabitty anisotropy. The test was analyzed with analytcel and numereal models. Permeabliy aniovopy was estimates to b6 26-1—to pencipal drectons of he peemcabity tensor ware not allgnad vith clsat rection, but apparenly were afecte by larger-scale ciscontinuties. Single-Phese interference-testinterpratation methods reed in prmesbilty estimates that could De used in the numerical analysis of multiphase testing during cavity competion provised ‘wellbore damage was icuded. The tects were condueted on the Nipon 0 Seam st Daws0 | Fiver, Bowen Basin, Ausvalis, Those high volatile bituminous coals are Pesan in 208 Prossure-bulldup tests wore performed! in this study in support of material balance ‘analysis of Frufland coal (San Juan asin) wel. Late-time Boundary effects were hnrprelee. OF parlculsr nota, ths study documented te possible functional form lexponeniily of absolute permeatiily growth as a function of deplation in Frutand coal wels—a permeability grow function derived from pressure-buldup tests was used in single-well numerical simulation modeling, This stay discusses the importance of injecting below fracture pressure fo obtain en curate analysis of injection alo tests, the impact of a nonstable reservar pressure, {and Importance of obtaining good communication with the reservot. Thay siess tho portance of estimating fracture cosure pressure before parferming the injection fallof test 50 that the tests can be designed not to ‘exceed frac preesure Recommendations on testing before the inocion allo test are give. ‘As with Zahnor (1867), this study quan fod absolule permeabilty growth in Fruitand coal wells using wall test analysis, The results were used 10 calibrate the Palmer- Mansood equation (1896) for predicting absolute permeabiiiy as funcion of depletion, “his. study quantified pareabity anisotropy using four well interference testing {conzal wel a injector, culer wols a6 cbsorvaten) and core anatsls. The magnitude of the permoabily valuce wore consistent Detween feld and core tests, but the ‘maximum and minimum perneabites obtained trom core tests were oriented at ‘approximately 80" fo the flerforenco test. The fests were conducted on coals i the ‘Darbrook Mine, inthe Syiney coal basin, Australi, “This study ciseussed wel testing techniques. used in the cry Horseshoe Goals of ‘Wester Canada. twas noted tat ila water rjucton fall tests Indicated very low permeatilly Tor the coals, yeking very dscauraging results, which was attributed 10 ‘he waiersenstive nature ofthe coals. They also noted thatthe stress-senstive nature of the coals and non-Darey flow effete during injection caused injootan tosts wih hiogen to. yl “affcut and nonunique interpretations.” Conventional ‘rowdowndbuldup teste were, therfore, cuggesto, in which individual coal seams ‘ne muliseam coal package were loltod and taste. ‘Well testing was used to develop = method for predicting porosity and permestilty changes as function of facture pore pressure and gas content and composition in ‘suppor of enhance d CEM [ECEN) recovery and COs sequestration projects. Geshart eta. (2007) performed time-lapse pressure transient analysis of flowbulldup fests, gatherad for walls in the Fruitland coal of the San Juan basin, to quantly ‘bsoiule permeabilty (eslinated fom otfectve permeability using relative permeabilly ‘euves and uid saturation eatmatas) changes as @functon of dapiouon. As withthe Zaher (1987), te functional form of the permesoiy qrourh was noted 10 be ‘omponenil. The study analyzed wala rom wih and extemal tothe profi Franc (cal Faray. Clarison (2009) analyzed single-scam pressure buildup data gathered from the diy Horeashoe Coals of Woetorn Canada, Toet reeuite used to canetrain inputs in for ‘ulllgyer simuiion and producion date analyst. Various reservir heterogeneities ‘wore inorpreted fom pressure-tuldup dervaives inluding multiayer, composite, anc ual porasiy bohavio: [Authors performed postsrac testing to eveluste the eleciveness of various etnuation freaiments fora Manrvile coal we, Zaher 1907 Feu Hopiins etal, 1988 ror aver and ‘vaughn 1998 sT,FBU Wold and setrey 1000 ntarferonco, IFoT Bostimetal, 2005. FOT, FaU Mavor and Gunter 2006 IFT (ges ane wet), PAU Giethad etal, 2007 Fou ‘harkson Fav Hyland etal 2010 Feu tise esr was recently discussed by Basan eta (2008) $e Clarkson (2009), ang te Hereshoe Canyon play in sete Canada as an example. inthe complex mica Horeahoe Can- yon coal, opesrs wil fen periodically rn spinner surveys fore isolating the con! seas or ones wi rie-pg stem ‘ies and peers with tandem gouges hong, beneath them. The Spline srveys aren 6 sn halle of preecton back {the individual scams or os! zoes; the indivi ronet-eam Duty are analy fom eervoir resin 2, ad skit “Te impect of nonsatic abel permealiy on the wel tes stegy and analysis has also Been edessed in dal in Fetromy 2011 SPE Resenoir Evaluation & Engineering recent history, Absolute permeability in coal is known fo change ‘during primary depletion operations, either because of changes in effective stres, Which act to reduce permeability, or because of ‘desorption (eotmenly refered to as "matrix shrinkage"). which fois fo increase permeability during depletion, Which effect bas the dominant contol an permeability change depends upon such factors as the level of depletion ofthe coal, saturation level (above ‘or below desorption pressure), sorption characteristies, gs Com= positon, coal mechanical properties, coal fabric, and in-situ stress ‘conditions, Inthe past decade, PTA methods have been sed t0 ‘quantify absolute-permesbility changes in coal duving depletion a soo 750~«7OD~=*CDSCSC« RReservolrProssure, pala 1000 5,368 1E +020 £54060 109 Permeability md -s02e “A SCO4E-004 2089E+026 a oA 506008400 [aia ami Gn tn Pry — Ean ain Poti) — Br toi om] Fig, 5-Example of estimated (effective and absolute) permeabilities as ¢ function reservar prassuro obtalnad om time-lapse Fo eran oat wll {Used with permission irom Roger Glerhart). Eitective permeabilities (lower curve) obtained tram the Fo 2 Fre convesed fo abgolute pormeabillles (upper curve) using estimeted saturations end relative permeability curves {Glsthar et at 2007), Note that tho tend inthe data hes been sssurred to be exponentiat with pressure, ‘Mavar and Vaughn (1998) observed shat well-test data covld be fused to quaniify permeability gains because of matrix-shinkage tflcts in the Puitand conl of tho San Suan basin; absolute pet Inesblity estimated from DSTs was compared 10 estimates from ‘hui tess performed later inthe life ofthe wells. More recently, CGicthast et al. (2007) performed time-lapse PTA. on Froitland coal wells within and outside the Fruitland coal fairway in he ‘San Juan basin (Colorado postion), multiple sbutinfouildup tests ‘wore performed during the production life of a well, resulting i Imultpie estimates of perseabiity which could then be ploted vs reservoir resmare also estimated from each test. These time-lapse PTS were then analyzed for functional form, whick bas been sed 19 yalidate analytical odes developed to predict perme- ability changes in coal (eg. Palmer et al. (2007), Clarkson ral, (2D10)}. Example results from a dime-lapse PTA are given in Fig 5. Tn adition to quantiying permeability changes during deple~ tioa, Mavor and Ginter (2006) used well testing f0 develop a method for predicting porosity and permeability changes 2s a fimetion of fractare pore pressure and gas content and composition ja suppoct of enhanced-CBM (ECBM) -recovery and CO,-seques- tration projets, The epproach used was to (I) measure the gas composition and intial permeability of the reservoir, preferably Using 4 production test followed by shutin text, 2) estimate the mount of pressure indices strain using an injeetion test with ‘watee and/or 2 weakly adsorbing gas (@g.,.N,); end (3) estimate the amount of sorpion-induced stain using an injeetion test with ‘strongly adsorbing gas. One of the unique aspects ofthis stady is the atempt to separate relative permeability effests from absolute permeabihity changes. 2 Permeability Derived Prom Production Dara. A relative newcomer to the teebnigues available for permeability estimation Of coal, type-curve and straight-line methods (anelogous to these ‘methods used for PTA) have been used recently for analyzing dry (Gingle-phase gus) and to-phase (gas and water) CBM reser irs, ‘Use of these teetiniques is subject tothe samo (reervoir-elated) ‘complications a5 well-est analysis; a summary of the historieal ‘work oa CBM production analysis wes provided by Clarkson {tal 2007) and Hs not repeated here. “As discussed in Clarkson et al. (2008), advanced produetion- snalysis techniques for CBM reservoirs cin be clessfied as: (1) type curve methods, 2) stxight-ine methods, (3) analytical and 6 ‘numesical simetation, andl (4) empirical methods. In this section, ‘ve diseass only the first rw metho ‘Type curve methods involve matching production date (aor malized for back presse variations) to dimensionless analyti- allempirieal solutions —zeservoir and simulation properties ean be extacted from the match, For CBM, the dimensioalest-vaiable tlefnitions require pseudovarables that account for both pressure hd saturation dependent zesorvoir properties (plus desorption) fand require soperposition time functions to account for varicble fateflowing presse data, Use of these modified varisbles, in combination with dimensionless variables develope for conven tonal resecoirs, bas allowed CBM production/lowing-pressure Gata to be analyzed ina fashion analogous to that for conventional reservoirs, Both single-phase (gas) end two-phase (ges and wats) ‘CBM wells have been apalyzed using this approach, eceuse to ‘hace CBM reservoirs require an estimate of relative permeability, ‘Clarkson etal. (2008) recommended that type-curve analysis Be performed in parallel with analytical and numerical simalation uch that the output from siawlation history metehing may be ted for pseudovariable calculation duting type-curee matching ‘Apreement between rype-curve and simulaion-derived perme~ Ability end stimlaton values should provide for more-unique tstimates, Single-phase (gas or wate) analyses generally are “hore straightforward except in cases where nonstatic permeability tfects our or where reservoir heterogeneities (such ax multilayer effets) exist ‘Straightline techniques for production analysis are analogous to those used in PTA; ance the corect flow regime is identified tsing derivatives or other diagnostic tools, data corresponding to that flow regime are plotted on a specialized plot designed to Iinearize the data set—the yas variable is typically rate-normal- ized pressure or pseadopressire drop, and the a-axis if saperpo- sition time oF rianalized cumolative production. For transient ‘low analysis, pormeability ean be uniquely determined only if ‘tradial-flow period is appacent, from which pemmeabibty con be ‘Extracted ftom the radial-flow plot (Hager and Jones 2001). For ‘oundary- dominated flowy, permeability may be extracted from the ‘lowing taterinl-balance pet if skin is known. To dae, tansient- flow sizightlne techniques have been applied only t0 dry coal ‘The flowing mateial-balanee technique, which is applicable to bounary-dominated flow only, ean be used for single-phase- and ‘wo-paseflow CBM reservoirs (Clarkson etal. 2008) ‘February 2011 SPE Reservoir Evaluation & Engines 2.9 OBI Forecast Gos Rate, Mee fo ow mwa Time, Days Acton + Metusew Pratino-Blaingeme Type-Curvos Radial Flow Plot ] I _|— 1 i eather ‘Superposition Time 2.P CBN Flowing Material Balance ‘mor got amt on 1 0 Noxmalized Cumulative Production Mitsct Fig, 6—Match of Uinta besin CBM-woll gas- and waterrate data using (a) analytical simulation model and gas-rate dsta on the fasiabtiow plot (0), Pratikno-Blasingame type curves (c), and flowing-materab-balance plot (a. Modified from Clarkson et al (e000), ‘An example of the use of advanced production-analyss meth ‘us for permeability estimation in coal is given next Fig. 6). Tn this two-phase analysis example {dry CBM examples are given in CCacicson (2009) end! Gerami etal, 2007), an integrated analysis, approach is aplied in which staight-line and typescurve methods ‘r¢ used in peraliel; but inthis case, the ssturation- and pessure- ‘dependent terms obtained from simulation matching (Fig. 6a) are ‘wed to caleulate the pseudovarabies in the dimensionless terms {or the type curves (Fig. 6e) and relative penmeability for straight line analysis (Figs. 6b und 6). This approach was suggested by ‘Clarkson et al. 2008) and is Believed to yield a more unique analysis. The permeability obtined from simulation and from Iype-curve and radialfow analyses are in agreement; the skin values calculated from these meihods were then used to calculate permeability fiom Mowing-matesial-balance analysis. AS with dry ‘oat analysis, derivative tetiques (nat show here) can be used to idemify flow ragimes. This example difers from the original ‘Clarkson etal. 2008) because of reinitaization of the dst set a appeoximaely 30 days because of an operstional chenge—tent Uinliaton rests in estimates of permeability similar (othe orisi- ‘al analysis, Famer, transient (radial flow) straight-line analysis for this well was not shown in that work. In the preceding example, revorvoir heterogeneities such as ‘aultlayer effects and dusl-porosity behavior were not encoun (ered. Daal-porosity effects Have not been observed in production Signatures for CBM reservoirs analyzed by the authors 10 date, rlilayer effects are Known t0 occur in some CBM reservoirs, ‘nd Clarkson (2009) suggests procedure for analyzing such res- eri, inluding a layer flowing-maleral-blance method. no Static permeability effects occur, the eype-curve and straight-line ‘methods may be altered to account for these effets, as discussed in Claukson etal. (2009), The inclusion uf mumechansin flaw Febemy 2011 SPE Reserve Fvluation & Erginoering into type-curve madels has been investigated recently. Building on ‘earlier work by Mohaghegh and Brtekin (1991), who developed type curves for two-phase single-porosity CBM reseroits, and dual-porosity type curves for nararally (ractured reservirs froma De Prat et al (1981), Thararoop (2010) developed CBM type ceurves that account for two-phase flow, desorption, and dual- porosity characteristics. The curves were developed using a ew rmoltimechanisic (Jal-porosicy, doal-permeability) mumerical CBM simlator, using a constan-Alowing.pressure constraint and assuming radial Tow, Type-carve sets were generated for various ‘combinations of initial fracture and matrix water saturation and reservoir presse, reservoir size, and slorativity. An example set is given in Pig 7. The oye curves were tested spains simulated ses only; the practical application of these curves for real field data is yet to be determined, ‘Aminian et al. 2004, 2005) presented CBM (gas and water) ‘ype eucves based upon new definitions for dimensionless rate and dime. The ype eirves were generated from CBM-numerical- ‘Simolatorovpat As with the Thararoop type curves, the practical application ofthe Anminian etal 200d, 2008) type curves to field data is yet to be demonstrated, Pradustionsanalysis methods have been used by the authors 0 ‘estimate pemeubiliy ac well Jocations that are then mapped for tse asa stating point for field reservoir simulation, and 10 tofy permeability changes as a fanetion of depletion ina fashion analogous tothe time-lapse PTA methods discussed cates, Reservoirsimlation history matching is a form of avanced production data analysis, and it may also be used to estimate solire permeability in coal, provided robust estimates of osher reservoir end stimulation properties are avslable. Our prefer~ fence is fo obtain estimates of absolute permeability using PTA or production analysis 4s a starting point in Delé simulation; the r 20 Fig. 7—Gas (a]and water (o) production type curve forthe system with « = 0.00%, p/P, 0:4, Xs Yo = 7.500, Sir, = 80%, sr diferent Spar From Thararoop (2010). ‘permeability valucs, or the tend in these values with depfetion, are ‘hen modified to fine tune matches to well dynamic data. Also worthy of mention for field determination of coal pet= tmeability are petrophysical logs and geophysical methods. For example, a method has been developed to detect fractured coals, and areas of enhanced permeability by using conventional log suites to calculate movable-fluid volume in the col cleat system (Rozak tal. 2002) Image logs, such as FMI, can provide evidence ‘of fracturing from which permeability levels ean be infered, but & standard approach for direct estimation of pecmeabilicy fram well logs remains cluive. Recently, Eneans (HYland et al. 2010) pre- sented an integrated method in which geophysical methods-—epe- cifically seismic-attribute analysis such as time structure, Gaussian, curvature, and “ant tracking” techniques —were used to identify areas of enhanced permeability in coal, ‘The results were comrelated with well-log and production data; this sty illustrates the potential for eeismie methods to be used 'o predict regions of enhanced fracturing, Relative Permeability Estimation. As with absolute-pesme- biiy estimation, relative permeability in coal has been quantified from core analysis, production analysis, and simulation history ‘matching. Unfortunately, far less work has been conducted 10 measure relative permeability in coal in the past 20 years, despite the important contol relative permeability bas on CBM production ‘charctersis. Ham and Kantzas (2008) recently summarized the findings of relative permeability stues in coals nine studies since 1973 were cited. The reason for the lak of core-based measure ‘ments of relative permeability is related othe dificult ofthe met surements and the difficulty in obtaining and preserving samples ‘Mavor (1996) summarized some core-hased studies performed up to that point and discussed some ofthe nesnces of measurement and reporting. Mavor aso summarized tho prosedures associated with steady- and unsteady-state measurements. The properties of ‘coal that make relative permeability measutemeat difficult include (Hama and Kantzas 2008) fable and heterogeneous nature of coal, Which makes it diffient to obtain a representative samples low porosity of facture network: suess dependence of permeability, ‘nd sorption of gases onto coal The following focuses on atempts to derive relative permeability in coal from field data. ‘Young and Paul (1993) derived relative permeability curves from field data using fallfeld simulation. As described in Paul (1996), “simulation experience tas shown that pseudorelative per eabilty curves developed for coarse layering in eoal ate steepor ‘nd mote linear than Iaborstory derived eaves" Paul goes on tO slate, “A more linear pseudoreative pecmeabilty curve reflects better developed cleats and higher vertical permeability, which ‘ae caused by Tow stress. Such coals havo a potential for greater _ges-water gravity segregation” Convray oa. (1995) note tht rea §ve pemeabilty curves derived from reservoir-simmlaton history ‘marching are problematic because all other reservoir properties entered into the simulation (absolute horizontal and vertical perme ability, porosity), as well as reservoir desertion and near-wellbore characterization, must be accurate. They firher noted inconsistent sesults in asolute-permeabiltyestimstion from well test if rela- tive permeability curves derived from another formation?basin acd 0 Sar=080, Su=020, 6 w Simulation history matching were used in the analysis. Conway “13 ‘etal, (1995) highlight many of the issues with the satus of relative” permeability estimation in coal that do not appear 19 have besa addressed in the lteratare effectively in the last 15 yess, Clarkson eta. (2007) and more recently Clarkton et al. (2010) suggested that relative permeability for some coals ean be obtined from production analysis by solving the radial-flow equation to ~ 8 and water for effective permeability to those phases, then Sclecting a base permeability fin these studies, base peemeability was selected asthe effective permeability to gas ata point when ‘lative permeability effects were considered negligible (i. when the coal was dewatered]. Mobile water satration was determined ‘rom water material balance. The approach (with some proceso seat E02 1-09. ‘% 10% 20% 50% 40% via {oimprove the uniqueness ofthe analysis resalts. For example, the type-curve match shown in Fig. 6 may be used to estimate frae~ fore balflength ind conductivity, Tn that examol, the hnydrulic facture was assumed to be infinite conductivity, and a very short fseturebalf-length was desived fiom the analysis i's possible m some CBM development seerarios that hydrau- siecinetre properties or wellbore skin may not he constant with production, For example, in the Horsesboe Canyon CEM play of Fruitland Gol Gas/Mater Relative Pormeability 50% 60% 70% SO 90% eee? 8 8 Vil Fig. 6—Derivation of relative permeabilty eurves for CBI reservoirs. Source: McGovern (2007). western Canad, the reservoir has been shown to exhibit primarily {hy (gas only) production [see Bastion etal. (2005), yet in our experience, some wells exhibit lat or even inclining production (ig, 9a), which some have suggested is related to the cleanup ‘of ciilling ftids that invade the near-vellbore region. Ifthe non Salc-skin effect ie aecounted for, ie production-analysis methods ‘escibed earier may be sed (Figs. S¢ and. 94). Tn this example the final “cleaned up” skin (0.6) was infered from a buildup test a Nomaized Rate, selDipsiice dot wor oom ear ont ) a February 2011 SPE Reservoir Hveluation & Engineering 500 +1000 +500 Normalized Cumulative Production, Mtisct Fig. ¢-Match of Horsoshoe Canyon CBM well exhibiting nonstatic skin using analytical simulation model to match gas rate and ‘Clinclative gus produclion (a and b), gasrate data on tho Fetkovich lype curves (c), and flowing-materiat-balance plot (2). Wolt Performance Monitoring Frultland Coal Fairway Producer Fruitland Cot Ate At 400! cH F000 = $300 oe a {al 200 Pa 1.00 7 10 209 900 «00 co 600 700, of F aor o Reserv recs (io) 3 Month [= Proicod > Mirinin Uk «Cue + Haronower ] Fig, 10—Example applications of survellnce data including the flolont,etfecive horsepower useage (compression efficiency), en {gas sampling (b) to morifor gas-composition changes with dopl in backpreseure-coolticlent calculation (C=qiAP*) and In (b). fd used 40 anchor the Final skin; skin before this was adjusted fo achieve the match in Figs, Sa and 9. The well-testderived tbsolute permesblity was also honored. The changing skin wes imcorporaed ino the data-processing algorithms fr the type-curve match and fowing-rateriat-balance match, Th the past docede, horizontal wells (single and multilateral) hanve hecotne 2 popular method For exploitiag some CBM reser voir. Many of the wells completed have been operole, natural Completions—to estimate the effecive (contributing) wellbore Tengtt, welltest and advanced production-anslysis methods may be used, Clarkson et al. (2009) provides examples of the use of straight-line and type-curve methods for snalyzing dry and ewo- piliase (gas and ver) hoeizontal (noastimulates) CEM well Recently, some operators have bees using mullfactured horizon: tal wells to exploit low-permesblity CBM reservairs, whichis con- sistent with what is currently being done to develop shale-gas res trvois. To our knowledge, no examples ye appear in he literatare ofthe analysis of CBM wells exploited In this manner. We antic Dale that well-iost and production-analysis techniques curretly being developed for shale-gas reservoirs may be used to analyze CBM reservoirs exploited with rultifractured horizontal wells Discussion: Use of Surveillance ‘Techniques To Optimize Recovery In previous sections, we diseusted current methods wo extract res- etvoir-property, vell-completion, ad stimulation-ficieny infor tration from CBM wells, Modern surveillance methods are needed to optimize eld development fully. Examples of modern surveil- Toc methods that ave curently being used for CBM development include mouitoring of flowing pressures, line pressures, flow ates, and temperature using supervisory control and data acquisition (SCADA) on individial Wells, pressure observation wells for ‘ontinucus monitoring of coal-seam pressures; production logging ‘wed to trace coatebutions ftom individually stimulated 2ones in ‘retical wells completed in multiple coal seams; time-lapse PTA ised to provide estimates of reservoir pressure and changes in clfetive permeability (gas andlor watet) during depletion; and ‘wellhead gas sampling to monitor gxs-compostional changes wing to relative adsorption effects (Clarkson etal. 2008). Sucvei- Tance data shown below aro measured peviodically or entiauously throughout the well fe, and do not inclade other entcal data ust ally measured only once during tke wel life, such ab gas contents, Sorption isotherms, well logs, inital pressure-transieat testing, and >neffactorestimalation testing, An example of the application of| over ofthese surveillance techniques is given in Fig. 10, n O vinase Cenrined oohars use of SCADA data to monitor changes In backpressure coef. 1é minimum-‘iting-rate calculations (a) and the use of wellhead jetion (McGovern 2007), Pressufe observation wolls were used ‘Additional surveillance dita that have been used recently include sutface deformation monitoring both tmeter and interferometric Synthetic apertare radar (InSAR) have beea used recenly (Kopema ral. 2009) in an etter to detect subsurface woverent of CO, as ppt of a CO,-ECBM/storae pilot inthe Pump Canyon area of the Sen Juan basin. Both itmeter and InSAR can detect surface defor. ‘mation, wich can be used to infer activity in the subswriace—for ‘cxample, monitoring ofthe CO, plume duning injection operstiors. Although tiltmeters have traditionally been applied im the oil and fas industy to map hyéranlic fractures, tho recent application of the technology for tacking Mbid movement could be quite wsefol fer both primary (CBM) and ECBMCrecovery operations. Another surveillance metiod worthy of pote, with potential increased apt tation to CBM, is microeismic monitoring of hydranlic-ractring trestments, Because ofthe dificult in predicting hydraulic-frackare [geometry «prion ubing existing hydsaulic-racure models, it is fntcipated that miccsoisnic monitoring of CBM well ractore treatments will increase, particulerly if se of mulifraccared-hor: ‘ontal-well technology inereases. ‘With complexity of completion and stimulation methods for ‘CBM exploitation expected to inerease, there will be an inrease Jn demand for cost-effective surveillance technology to essst with wwell-performance monitoring and effective field development in primary- and exhanced-zecovery scenarios. Conclusions [After several decades of exploitation of CBM, much fas been feared about how to characterize this reservoir type, although Clearly thee fs stil toom for improvement is certain areas. The furrent work has summarized advancements in the assessment Of key reservoir properties such as gas content and permeability, although not all opics were covered. For example, geomechanical property assessment, erica! for hydraulicfractare and wellbore eametry design, was not covered in detail the reader is refered to recent york [e-g., Deisman etal (2008)] on this subject. “The fature of CBM exploitation will include development of deeper, lower-quality CBM reservoirs, necessitating use of ‘advanced ‘dailling and completion methods. As a result, new procedures for reservoir ovsluation will need to be advanced to Improve, for example, interproation of pressure-ate-iansient Signatures associated with complex wellbore/competion designs “Aithough it is expected that CEM. dovelopment vill continue t ‘benefit rom advancements i fechnology related to exploitation of other unconventional reservoirs such as shale and tight gas, CBM- specific technology will need to he developed, particularly when it comes (0 implementation of eahanced-reeovery operations. ebony 2011 SPE Refervole valuation & Engineering Nomenclature A = dminage ares, acres 2, = gas formation volume facto, resesvoir volume to surface Yoluane G, = gas compressibility, Pa"! (Ea. 9) FF) = gas desorption ratio, cumulative gas desorbed t total gas ‘esarbed (Eq. 8) FF = gas residual rato, remaining gas to foal ane to be taken up by sample Fy = gasuplake ratio G, = original gas in place, Bsc (Eq, 4) G, = cumulative gas production, Meet or Bact 1 = formation thickness, 4 = absolute (formation) permeability, md or m? (Eq. 9 and Fy.) = pressure, psia by = Langmuir pressure constant, psia AK, = radius of dsil cores, m (Ea, 9) 5, = slope of straightine ft to Fy vs. squsre root of time at cay time, dimensionless (Eqs. 10 and 1) 45, = slope of strapbt-line i 10 1n(¥4) vs, tiene at late time, dimensionless (Eqs. 8 and 9) 5, = water saturation, dimensionless, fraction 1 time, days or seconds (Bqs. & and 10) ¥, = Langmuir volume constant, sf/ton 2 = mocified compressibility factor, dimensionless iy = 28s viscosity, cp or Pas (Ege. 9 ad 11) = porosity, fraction = facture storage facor, fraction P. = coal bulk density, glen? 4; = fit roo of Bessel equation J, €) =O, dimensionless Eq, 9) Subseripts D = dimensionless decline variable gE 7 initial R= reservoir se = standard conditions W = water Superseripts altered variable average property Reforences ‘Abed, Celle, A anil Rowx, B. 2006. A Genersied Matta Balance Bgvaton for Crlbed Methane Reserv. Paper SPB 102638 Presented a the SPE Anal Tesical Confsence and Exhibition, Sen “Acton, Tera, USA, 24-77 September oj 1021181025384. ALKhalfoh, A-TA, Hore, Ri, ad Aza, K 197. Ib Pls Determine ‘ion of Reserve Relative Pecmeablity Using Wel Tet Anaya, Paper SSPE 16774 presened a he SPH Anowal Tectaieal Conference and nibitor, Dalla, 27-30 Sepestbor. doi: 102118/16774-M5. Arminia, K, Ame, 8, Bhawar, A, Sanchce, Mend Gata, A. 2004, ‘ype Curves for Cotbed Methane Production Prediction Pages SPE 91482 presented at he SPE Eastern Regional Messe, Chrkstom, ‘ies Vigna USA, 15-17 September di. 1021189148208. Alain, K.. Amed S, Bhavss, AB. abd Latshminarayano, S, 2005 ‘ype Curves for Prodacton Prediction end Evaluation of Coubed Methane Rescirs. Paper SPE 97957 presented at he SPB Fata Regional Mecing, Margao, West ign, USA, 14-16 Septem ber doi: 10.2118197059.M8. AS 2980/19, Guide To The Determination Of Desorbable Got Content OF Cont Seams, 191. Nach Sydney, Absa: SaodardsAssOiton of Asta ‘anc, Rame, VL and Cris, DP-2007 Hott Fractare Diagnosis, ‘ape STW 1077 peste te Rocky Mowat Oi ang Gs Technol ny Symposium, Denver, 16-18 Apa doi: 102118787 MS © Retnuary 2011 SPE Reseroir Brolin & Erginserng Bastion, PA.Wird, OR, Wing Land Von, GW, 200, Assessment ‘and Development of te Dry Horseshoe Canyon CBM Play in Cann, Piper SPE 96909 peotened atthe SPE Annual Techical Conference sn Exibition, Dallas, 9-12 October. di 10211826899. Bell G. 2005. Gat Conte! Varition of Horoshae Canyon Coals An no- Iyial Review of Masument Reval resented athe CSPO Gussow Confrence on Coalbed Metaana, Canmore, Cand, 9-11 Misch, ‘Boule. and Walls. 1982, Pulse Decry Permeability. ty Soh Vion and Experimental Tet SPE J. 2 (3): 719-22. SPB.SPAA-PA, do 1o2iiE974a-PA. Brace, WE, Walsh, 1B. and Frangos, WT: 1968. Permeshilty of prsite ‘ander high presi.) Geophys es. 73 (6) 2225-2236, do 10.1025/ mBorson6p02225. Dustin, AAD. and Bustin, RM, 2009, Gat fa Box: How Much Frodac- ible Gas is inthe Horseshoe Canyos. Report, Canséaa Society for Uncoovetional Gas (CSUG), Calgary (Notember 2006) Basin, AMM, Bustin, RM, and Col, X. 2008. lmpaet of Shale Face on Production. Paper SPE 14167 presented atte SPE Unconventional Reservoits Conference, Keystone, Colorado, USA, 10-12 Febraay doe 10.2118/18167-M8, Bustin, ROM. 1997. Importance of ubsc and composition on the sess sensitivity of permeability in some coals, notern Sydney Besa, Asia; clevanc to cclted methane exletaton. AAPG Bein 81 (1): 1894-1908 Bustin, RM. 2005. Coalbed Methane Geology and Engineering: Appliea- onto Exploration and Develepoent. Calgay: CBM Solutions, Bustin, RM, and Clarion, CR. 1990. Free Gat in Matrix Poros: & Potentially Substantial Resource in Low Rank Coat. Proc, Lemna ‘ons Coalbed Methane Symposium, Tuscaloots, Alabim, USA, 3-7 May, 197-916 (Clarkson, CR. 2009, Case Say: Production Daa and Pressure Tiansent, Analysis of Horseshoe Canyon CBM Walls. J) Con Pt esol 8 (Oy: 27-38, SPE-114485-PA. de 1021 18024585-FA (Cason, CR snd Busi, RIM, 2010. Coalbod Methane: Current Era ‘ion Method, Fue ‘Techical Challenges. Papee SPE 131781 pe ‘ened atthe SPE: Unconventional Gas Confereace, Ptaharh, Pen syvania, USA, 21-25 February. doi 10.2018/131791-MS, (Garkson, CR. and MeGovera, JM, 2001. Stay ofthe Potente Impact ‘of Matrix Free Gas Storage Upon Cate Gas Reserves snd Prod sion Using a New Mateil Balance Equation. PaperO113 presented a ‘he Inemasinal Coslbed Metiane Sympostm, Testis, Alban, USA, 14-18 May. Clarion, CR and’ MeGovem, IM. 2005. Optimization of Coslbed. Mathine-Resevois Eaplration and Developsat Stseies Trough Inegration of Simulation and Economies, SPE Res Eva & Ene 8 (8): ‘82-519, SPE-BRBA3-PA. dt: 10211888843 PA. (Chnkson, CR, Jorten, CL, Geran, RR, aed eile, 1.2 2007 Prods io Bata Aealyis of CBM Wel, Paper SPE 107705 pest the Rocky Mounisin Oi aod Gas Techelogy Ssposin, Denver, 16-18 ‘Apa: dis 1021187107705-MS. (Clarkson, CA, Frdun, CL, ier, Rk aul Seidl, JP. 2008, Prcace tion Data Analysis of Coola Methee Wells SPE Res Ev & Bg Mt (Q}; 311-325, SPE-UTIOS-PA, dois 102018107705. ‘Clarkson, CR. Jorden, CL, Th, D, and Blinn, A. 2009, Princ tion Data Analysis fractured ed Horizontal CBM Wells Paper SPS 125629 presented atthe SPE Essern Regonal Meeting, Chireston, ‘est Vigna, USA, 23-25 September. du 10.2118/129929-M5, (Clakson, CR, Fan, Z, Pala, 1, and Harpalani, S. 2010. Predicting Sorpion-InueedSiain aad Pecmeabliy Increase With Depletion for (Coals Methane Reseevis.SP& J. 1S (I 152-159, SPE1}4778 PA do: 102118114778 PA, Carson, CR, Raimsaian, Mend Kans, A. 2010. Relative Penne- bility of CEM Reservoirs Paper CSICISPE 137404 reseed nt the Canadian Unconventional Resources nd International Petroleum Conference, Calgary, 19-21 Ocxnber: do: 10211871 STAM. Conway MI, Mavor, ML, Sausberry, J, Bare, RB, and Selvafonge, RA. 1995, Molt:Plure Flow Propecies for Calted Metbane Wells ‘A Laboratory an Field Stuy. Psper SPE 29576 presented a the Low Pemeailty Reservoirs Symposium, Denver, 19-22 March, do 10211829576. Cox, DO, Youvg, GC. and Bell, MJ, 1995. Wel Texting in Coatbes ‘Metane (CBM) Well: An Environments! Remedition Case. Paper B { ‘SPE 0578 presented at the SPE Anal Techalesl Conference ant Exhibition, Dils, 22-25 Octobe. doi: 10:21180578.MS. (Gai, X and Bustin, RM. 2005. Contos of Coal Furic on Cotes Gas Producion and Composition Shit ia Bola Felé Production and Canister Desorption Test. SPE J. 11 (1): 11-119, SPES5035-2A. hi 10-2118785035-PA. (Cu, X, Bustin A.MIM. ond Bas, RM. 2009, Measoremenis of gas pes “meaty ad dtfsvc of ight resecvce rocks: diferent approaches fad thet applistions, Gee 9 (3) 208-723, do: 10.1111.1468- 123.200.0244. Cussler, EL. 1997. Difsion: Mase Transfer in Phd Systems, second ‘eligon, Cambdgs, UK: Cambeidge Seis in Chemical Enginesxing. (Contidge University Pres Da Prat, CinooLey nd Ramey, H Je 198]. Decline Curve Analysis ‘Using Type Cares fr Two-Porosity Systms. SPE 1.21 (3): 354-362, SPE-O292-PA doi: 1021189292-PA. Delman, N, Gents, and Chaletinyk, RJ. 2008, Unconventional ‘romechinial esting on coal fr coalbed reseroie well eign: Tae ‘Aber Footils and Plains. Ineratonal of Cool Geology 78 (0) 15-26 dls 101016 coal 2007 12.004, Diamond, WP. and Schaal, $1. 1998. Measuring the gas content of onl: A roview,nemationn of Coal Gelzgy 38 (1-4): 311-331 foi 10 101680168. 5162¢97)00040-2 Dicker A and Sma RM. 1988, A Prasat Approach for Determining ‘Pernoibilty From Laboratory Presse Pulse Decay Mersaremens. Paper SPS 17578 presntd tthe Icaational Mesing in Protea Egincsing, min, China, 1-4 November. doi: 10.2118/17578-MS. Feikovich, MD. Guero, 7, Feevic, My and Thomas 1K. 1986, Olan es Relative Pesmesiniies rom Rate Te Performance Dat Piper SPE 15431 presented atthe SPE Annual Technical Coneseace and Exhibition, New Orleans, 5-8 October. do: 102118/13431-MS, etkovich, M3. 1980. Deoine Cove Aaalyss Using Type Curves. J. Per Technol 32 (61065-1077. SPE 4629. di: 102118H4620-. Geran $, Pool-Darvish, M, Mond, K. and Mata, L. 2007. ype ‘Curves for Dry CBM Reserve Wis Exuilbium Desorption. Paper 1207-011 presented the Petroeum Society's CIPC/Aanual Techneal ‘Mesing, Clery 12-14 June Ciera, IER, Claiton, CR, abd Seite, LP 2007. Spatial Vaviaion of| ‘San fon Dain Frttand Cabo Methane Pressure DepencntPeme- ily: Mogitae ond Puntionat Form. Paper TPTC 11333 presente ft the Inerational Pevolaam ‘Technology Canterence, Dobe 4-6 ecenuber dor 102523/1353.M6. age, CJ, td Joes, LR. 2001. Analyzing Flowing Prodoeton Daa With “Standard Pressure Transient Methods. Paper SPE 71033 preset at the SPE Rocky Mouttein Petloum Tesinlosy Confeence, Key: stove, Colorado, USA, 21-23 May, do 10:211871023-MS, ‘lan, and Kanzas, A 2008, Mesorement of Relative Pornelity of ‘Cou Approaches ad Limitations. Paper SFE. 11494 presented at the CIPCSE Gas Technology Symposives 2008 Joint Conference, Calgary 46-19 June. doi: 102118/118004-M5. Hoploes, CW, Ponts, LE, Fumorel RW, and Spivey. LP. 1998 tals ofInjecton/Fllet Testing ln Coalbed Methane Reservoirs. Paper SP 39772 presested wt the SPE Pecrian Basin Oi) and Gay Recovery Convene, Midlacd, Texas, USA, 25-27 March di io2iisao7m-Ms. oie, PA. Tracy. LV, Neva, CE, Bredehoe J.D, nd Silliman, SF SRL. A trie iaberetory method for detaining the tara ‘ropes of tight rocks—L Thoary. Intemational. of Rock Mec Tes nd Mining Scenes & Geomechanics Abstracts 18 (3) 245-252, doi 10.0160 188-9062(81)50579.7, lod, F, Paltal,F, Yo, M1, Abaco, C., fnking, Ny and Cox, W. 2010, ‘Manavile CBM: laergrated Approach in the Mikwan Area Esper ‘SPE 138114 presened atthe Canadian Unoometionl Resources tn Intemational Petroear Conference, Calgary, 19-21 October. di 1021387138114 WS. Sense D. a Sith, LAK. 1997. A Pacieal Apps to Cotbed Mesh- fae Reserve Poison Using A Modified Maes Balance Telique Paper 9165 presets atthe Inecrtional Coase’ Merbane Symp umn, Taetooe, abun, USA, 12-7 May. Jochen, VA. abd Lee, WJ. 1983. Resevoir Charceization of a ‘Opeahole Canty Completion Using Paeducion and Well Test Data [Ammiyass Paper 26917 prevented the SPE Easiam Regional Confer ™ ence and Exhibition, Psburgh, Penosylvania, USA, 2-4 November, (doe 1021187209178, Jones, S.C. 1997. A Techaiqu foe FnterPuse-Decay Permeability Met ‘uemens in Tight Rocks, SPE Form Bval 12 (1): 19-26, SPE 28450. PA, dls 102118728450-PA, “Kamal, MIM ad Six, 1.1998. Presse Tnsient Analysis of Methane ‘Producing Coslbes. SPE Adsanced Tecnology Serie 1 (Ds 195-200, SPE-I9789-PA, do 102118/19789-PA, King, GR, 1990, Matai Balance Tetelgues for Coal Seam and Devo. an Sbale Gos Resovoie.Papee SPE 20730 prescutd atthe SPE ‘Annuol Tecnica Conference and Exibion, New Orleans, 23-26 Septem. di 10.2118720730-48, ing, OR. Estskia and Schwere, EC. 1986, Numeral Simuliton of ‘he Transient: Behave of Coal Seam Degsifcation Wells. SPE Form val @ 165-183, SPE-12259-PA, doy 10.2118/12258-P. Kissel, EN, MCalloch, CM, and Elder, CH. 1973, The Dieot Method ‘of Detrning Methane Content of Coelbeds fer Venilation Design Report of Uvestigaions 7267, US Burew of Minss, Washington, be. opera, Ge, Oatinot, AN, MeColpin, GR, Lit. N. Heth, JE. "Wells, A. and Yousg, GB, 2009, CO, RCBMStorage Actes a he ‘San Juan Basin’ Pump Canyon Test ie, Paper SPE 124002 presents athe SPE Annual Tstnical Coseonce and xhiiton, New Osten, 47 October, do: 1021 18/126002-08. Lamorre, Ru al Pope, J 2007. Ceteal-Gs- Conta Technolgy Pro- Wiles Couibed Metiue Reseroie Dats. J. Per Technol $9 (11 106-113. SPE-103520-6. ‘Langmaic 1 1916. Thecoataton ad endamectal properties of sli and igus ofthe American Chenical Society 98 (11): 2221-2298, Lubooh, SE, Manet, RA, Olson, IE and Soot, A.R. 1998, Charaters- ties an orgs af coal cea A review fterational J of Coat Geology 236 (1: 175-207, de 10.1016150166-516200700012-. Maver, MJ 1996 Coalbed Methane Reser Propertics. In A Guide te Cealbed Methane Reservolr Engineering, © 1. Saliboy, BS. Schafer, and RA. Schraufayel, Report GRL-94/0397. Chicago: Gas Research Inte ‘Mavs, MI and Gane, WEB, 2006, Soccer Porosity and Peomeailiy ‘of Coal vs. Gas Composition art Pressure. SPE fer Bua & Png 9 (2), 14125, SPE 90255-PA. dois 10211890055-PA. Mayor, MJ, aid Nelson, CR. 1997. Coalbed Reservoir Gas-o-Pace “Analysis: Report GRLTT0263, Gus Reach nsiute, Cheap ‘Mavor Md and Prt, J, 1996. knpoved Meldology for Deven "Total Ons Cone Volume I. Compare Evaluation ofthe Acwicay ‘a Gas-i-Phce Esnats and Review of Lost Gas Models. Topics! Report GRI-940429, Gas Resch Inst, Chicago (March 1995) aver, MJ, ad Robinson, TR. 1993. Analysis of Coal Gus Reservoir hterfrence ani Cavity Well Tes, Paper SPE 25860 prescott the Low Peteatlty Reservoirs Symposium, Denver 12-14 Apri. do 102118725860 M8. Mavor MJ, aud Saslsbony, J 1996. esting Coad Methane Wells 18 "A Guide to Goebed Methane Retervoir Bagineriag, Sauls bony BS, Schafer, snd LA. Schanfagel, Repor No. GRESAA997 Chicngo: Gas Rescarh Insti. ‘MavorMEJ end Vang, J 1998, erasing Coal Absoite Pemesity ‘San Jaa Basin Pruiand Formation SPE Res Boal & fing 1 201-208, SPE-AB105-PA. doi: 10.2118739105-2A, Maver, MLS, Close 1C, and Prat, TJ. 1991, Sumer ofthe Compeion ‘pimization and Assesment Laborato (COAL) Sie. Tiel Report (GRL9L0377, Gus Research Itt, Chicago (I December 199D, Movor, MJ, Pr, 1, aad Briton, RN. 1994, Improved Methodology for Deterining Teal Gas Content Vl. I Canister Gas Desorption Data Saniary. Topi! Report GRL-SMOSIO, Ons Rescarch Institue, Chicago (iy 1994) MeGover, JL 2007. How Is Betonie Measurement Being Utlzed? Presented at the SPE Applied Technology Workshop, Life of Field Surveillance in Unconventional Gas Resover, Santa Fe, New Mexico, USA, 31 July? August. ‘MeLzanan, J.D, Sehefer, PS. and Peat, TJ. 1995, A guide to dteanin- ing cuties gus content Publiston No. GRE 840306, Gas Reseach Insite, Chicago. Mohaghrah SD. andEtcin-T 1991.4 Type-Cuve Solution orCoal Seam ‘Beysifiation Wells Producing Under ‘Two Phase Flow Contos. [Febrtry 2011 SPE Reservoir Evaluation & Pgineeing Paper SPE 72673 presented atthe SPE Annual Teimieal Conference and Exhibition, Dallas, 6-9 October. di: 1021187226738. Primer, . 2008. Permesality changes in coal: Analytieal modeling Thieretional 1. of Coot Geology 77 (1-2). 119-126. dois 10.1016! | 2008 09.006. almes, 2010, Coaibed methane completions: A world view. Iter Thational J. of Coal Geology 82 Ged). 186-105. do 10.0160 jal 200012010 ales, ead Maori J, 199, Hom Peneabilty Depends on Stes nt Pore Pres in Coaltede!A New Med. Paper SPE 36737 presente at the SPE Annual Tetnical Conference and Exhbion, Denes, -9 October doi 10211886737. z Pines, 1, Lambert, SW, ard Sper, JL. 1983. Coabed Metane Well ‘Completions tnd Stimoltons. I Hydocarbons from Coa, et. BE Jaw and DID. Rice. Tls: AAPG Studies in Geology, AAPG. pales I, Miter, Mend Gunter, B. 2007 Peraeability Changs in Cos, ‘Sears Doring Produstion and Injection. Paper O713 presente atthe Imerrational Coated Methane Symposium, ‘Tusculoss, Albams, USA, 5-9 May i, G.W. 1996, Simslatng Coatbed Methane Reserves. In A Guide vo Coabed Methane Revervoir Buglvering, ef. 1. Sausber. PS. Schafer, and RA. Sebastnagel, Repost GRI-SAR9T. Chicago: Cas Resear Iai ‘erin, RL, 1956. Analysis of Pressure Buléup Curves. APY Dring ae Production Practice (1956): 482-309 Ranuthy, M, Masjenison, D.M., and Daves, SB. 2002. Diagnontic "cture Injection Test in Cons o Deters Pore Pressure and es- Imes. Paper SPE 7570) presented atthe SPE Gas Teennology Symposim, Calgary. 30 Apri 2 Ma. doi 1.211875701 M5, Rol, AT, usin, RM, Stash, 6.W, Beaton, A. Richardson, Rd ‘wer. 2002. Azlicaton of LogFAC® to Coated Methane Ilo tation in Wesera Canada: A Case His from Arley Coals nest Red Deer, Alberts, Paper SPE 75676 peseted atthe SPE Gas Tecnology Symposiom, Calgary 30 Ape!=2 May, do: 10.211875676-MS5, Ruckansei,E, Vakdyinathan, AS. ad Yaurgquist, OR, }9T1. Sorption by soiks with bidipece poe suvcanes. Chemical Engineering Se ences Yo 9): 1305-1318. ei 10.2016008.2509¢71)80051-9. Seiler, G, 2007 Using Werle Formation Evaation Tools To Char ‘size Couled Metre Formation. Paper SPE 111213 presented f the Ease Regional Meatng, Lexington, Kenney, USA, 17-18 (cso: do: 102138/111213-M8. Sc, LP 1999. A Modiedp2 Method for Coal Wells Paper SPE $8608 resend at the SPE Roeky Movatan Regional Mestng, Gillet, ‘Wyoming, USA, 15-18 May. do: 10-2118/55005-MS. Scie, TP, Sansone, Dl, and Evckson, DJ, 1992. Application of Mateus Geometry to Sues Depevea Penneabiity in Coals. Paper SPH 24361 presented at the SPE Rocky Moanin Regions! Meeting, ‘Cesar, yeming, USA, 12-21 May. do: 1.2118 24361-MS. Sella, LR, Kats, GM, and Kease LD. 1991, Pressure Fllolf Tests of ‘Now Coal Wells. Peper SPE1209 presen! the Rocky Monrtain Regional nd Low: Permeaity Reservoirs Symposinn, Denes, 1-17 Api do 10:211821800M5, Sho, DM, Lakehtsanan, CC, end White, N. 1994, Eaton ofa “Coal Penney fom Sing and Packer Test, Paper SPE 28664 all tle ftom SPE, Richardson, Texas, USA, ‘Smit, DIM, tnd Willams, FL. 1981. A New Technique for Datei ing the Methane Content of Coal, Proc, Intesocity Energy Con- version Engineering Conference, Ovando, Pirida, USA, Avgis. ner127. Spaiks, DP, MeLendon, TH, Savlsbeny, HL, and Laser, $.. 1995 ‘The Bites of Sisson Coal Reservoir Performance, Black War ioe Basin, US.A, Paper SPE SUT presented at he SPE Annal ‘Tehnical Conference and Exubion, Dallas, 22-25 October. do 10211930784 MS. ‘Terwoop, P2010. Development of a Multi Mectanistic, Deal Porosity. Dual Permesblry Numerical Flow Model for Coalbed Meane Febuary 2011 SPE Reservoir Bvauaion & Enginecting Reservoirs Accounting for Coal Snkage en Swelling Becks. PRD thers, The Pemsyvania State Universi, Department of Pregy ae ‘Minefl Bngnceing, College Tovaship, Pensylvana, USA (Angus 2010) “hima, D Bonner, B., Heasd, H.C, and Duba, A, 1980. Bit of Prose and Stetson Water Transport in Inet apd Fareed Gxb- te and Grants J of Geophyricel Research #8 (B12): 7059-7071, i: 101029120854 2907053 Waren, JE, tnd Root, Pd. 1963, The Bebavior of Noncally Frases Reservoirs, SPE 1. 3 (245-255; Tras, AIME, 228. SPEAZO-PA. thi 10 2118626. Wei, XR, Wang, OX, Massaroto, P, Golding, .D, and Ralph, V. “2007. A Review om Recent Advances inthe Numerical Simulation or Coalbed-Methane-Reoovery Process. PE Res Eval & Eng 10 (6) (657-666, SPE-S3IOI-PA. dot 102U18051D1-PA. “Wold, MB, and efrey, RG, 1989. A Comparison of Coal Seam Dirc~ tional Pecmeaiiy 8 Messured ie Laboratory Coce Tet and in Well Iaveerence Tests Paper SPE SSS9R presented a the SPE Rocky Mountain Regional Mesng, Giete, Wyoming, USA, 15-18 May. doi 10.211853598-m8. ‘Wold, M.B, Choi, SK, Koenig, RA. ad Davidson, SC. 1996. Anso- ‘won Seam Response to Two-Phase Fei lection Ito @ Colbed| ‘Methane Reseroir—Messoement and Simulation. Paper SPE 36984 pvesenet atthe SPE Asia Pete Oil and Gas Conference, Adelaide, ‘Australia 28-51 October. dois 102018036084-MS. ‘Yamata, SF and Toacs, AHL, 1980, A Revow of a Pulse Tecrigue for Pennesbiliy Menssremets. SPE J. 20 (Sj: 397-358. SPE-STOORA, ois 10.2118 760-2A 1¥3, Addl, LY, and Dense, C.¥, 2008, Gas Transport n Bidisperse ‘Coat Panicles:invesigaton for an Bective Diffsion Coctcent in {Goal Beds Can Pet Toto 47 (10): 20-26. FCPT Paper No. 08-10- 20, do 10.211808-10-20, ‘Young, G.B.C. eed Pcl, GM, 1993 Reservoir Charactesiaton of Mary ‘ee and Black Creek Cons tthe Rock Crack Feld Laboratory, lack Wirror Ban, Topical Report GRE-S3/0175, Gas Research Tstiue, Chicano (May-December 1992). “Zale, R. 1997. Appiestion of Material Balance co Determine Uiinwae [Recovery ofa San Jnxn Fruilnd Coal Well Paper SPE 38858 pre- sented at ne SPE Aamaal Technical Confarace and Edition, San Antoni, Texas, USA, §-8 Oecebe. di: JO2TIS/8858 MS. Zaer M.D. 1906, Basic Reservoir Bagnceing fr Coal. A Guide to Coat bed Aothone Reservoir Bploerng, Tpieal Report GR-9&NSIT, Cas Reese Insite, Chicago. ue, MD, Spats, DP, and Lee, WI, 1990, Design and Interpretation of injectionFall ess for Cased Methane Wells Paper SPE 20569 presented atthe SPE Ansval Technical Confrence ap) Exhibition, [New Orleans, 23-26 September, do 10.211820569.N8, Christopher R. Clarkson (PhD, PEng) Is a. professor anc the Encane Chal in Uncorwentional Gas in the Department of ‘Geoscience, Univaraiy oF Coker, and presdent of UOC ‘Conauting Ine, E-mail clerigoo@ueaigaryca. Before comings fo 2eagemnla, he werkod In indusry as a Teservoit engineer in Unconventional gos. Cerkson’s work focus was on exploration ‘or and development of coalbed mathane, shoie, ight gas, ‘ond conventional reservols. He has sludiad both prkreny and Ennonoed coubed mahane recovety processes in fhe San ‘an bash Sf New Maxico/Colorado and 's partculorly net “et in the physics 0 gos slorege and rongpor in unconwen fionel gaz reservoirs and he application of production date, ‘onahvse/revervelskmutation fo opticnize gas vecovary ond Wal performance, Clarison holds © FAD degree in geologicat fsnghneering fom the Unversity of Brith Columbia. He was on SPE Distinguished Locturor in 2009-2010. R. Mare Bustin Ph. (Goot H&8C) Ib0 protester of petroleum ancl coal geotogy in he Bapormant of Earth arid Qcean Sciences at the Univershy of Bien Courrbia ancl presklent of RMB Earin Science. Bustin ‘on slectad Felew af ihe Royal Society of Canada. 6

S-ar putea să vă placă și